LaBSE-en-ru-myv-v2 Open-source Multilingual Embedding Model - Supports Precise Applications in Russian and Moksha

Labse En Ru Myv V2

Developed by slone

This is a multilingual embedding model based on LaBSE, supporting Russian and Moksha languages, fine-tuned on the myv_ru_2022 dataset.

Downloads 59

Release Time : 11/1/2023

Model Overview

This model is a variant of LaBSE specifically optimized for Russian and Moksha, suitable for bilingual text mining and masked language model tasks.

Bilingual Support

Specifically optimized for Russian and Moksha languages

Performance Improvement

Outperforms v1 version in bilingual text mining and masked language model tasks

Pre-training + Fine-tuning

Based on LaBSE model, fine-tuned for approximately 150,000 steps on the myv_ru_2022 dataset

Text Embedding

Sentence Similarity Calculation

Masked Language Modeling

Feature Extraction

Natural Language Processing

Bilingual Text Mining

Used for text similarity analysis and mining between Russian and Moksha

Performance improved compared to v1 version

Language Model Tasks

Used for masked language modeling tasks

Performs well on Moksha language tasks

Property	Details
Language	Russian (ru), Erzya (myv)
Tags	erzya, mordovian, fill-mask, pretraining, embeddings, masked-lm, feature-extraction, sentence-similarity
License	cc-by-sa-4.0
Datasets	slone/myv_ru_2022

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base